Harnessing the hyperlink structure of the Web

The idea of hyperlinks came up basically as a navigational aid for browsing hypertext resources. But it has been very well harnessed for another very useful purpose - searching the hypertext repositories. In fact, this is arguably the best instance in which a thing is better utilized for a purpose other than what it was primarily meant for. We substantiate this claim in this paper, by highlighting the three main usage the hypertextual structure has been utilized for. We begin by showing how the hyperlinks are used for propelling the ranking schemes for web search. The notions of hubs and authorities on one hand, and PageRank on the other, are the keys to ranking the web search results. Next we take up the problem of finding pages related to a given seed page on the web. The motivation comes from the situation in which we don't know the precise name of the thing we are searching, rather we have a sample to be compared with. Again the hyperlinks structure comes handy in this regard. Finally, we take up the questions pointed at characterizing the web. We primarily consider the hyperlink based random walk schemes for the web, for not only comparing the quality of various search engines, but also getting to know more difficult attributes like the size of the web, the percentage of corn sites, the average size of a web page, etc.